Impact of Damage on Groundwater Flow Dynamics in a Compound Rockslide

The amount of internal deformation or damage created in a mature rockslide depends primarily on the basal rupture plane geometry and total amount of accumulated displacement. We present results from a 65 million m³ compound rockslide (Cerentino, Switzerland), which started to creep along a compound sliding surface about 5000 years ago. Investigations of the landslide body over the past 40 years include 8 deep boreholes, diverse monitoring systems, and geophysical as well as geomorphological investigations. The data set generated is unique and allows the quantitative linking of damage to hydrostratigraphy, groundwater recharge, and groundwater flow dynamics.   

The long-term creep of this crystalline rock landslide body along a stepped and bowl-shaped main rupture surface has led to a total displacement of about 500 m. Damage of the landslide body has been studied in great detail using a high quality triple tube core drilled in 2017 through the landslide body and into the stable bedrock down to 228 m depth. Inclinometer and fiber optic displacement measurements along this borehole suggest that the main sliding surface is located at 107 m and that significant distributed deformation occurs in the coarse-grained blocky carapace of the over-steepened landslide toe. In addition, several secondary sliding surfaces could be detected down to a depth of up to 207 m.

The landslide mass is heavily damaged and consists of variably broken cataclastic rock down to 140 m depth with grain sizes dominated by cobbles, gravel, sand and silt. From 140 to 170 m depth we observe a fractured rock mass with thinner kakirite sections. Below 170 m the rock mass quality is good in terms of RQD (40-90) and fracture density. 20 samples from cataclastic layers have been analyzed in detail with respect to grain size distribution, water content, and mechanical properties. Combining grain size analyses with a heating test conducted after borehole completion, we derive a detailed hydrostratigraphic profile through the entire landslide mass.

Groundwater discharge monitored at the landslide suggests high recharge rates for an alpine catchment (772 mm per year on average, or 0.7 Mm³), and can be balanced if we consider that there are no significant regional contributions from surrounding systems. Groundwater storage-discharge relationships were quantified based on spring recession analysis and a simple rainfall-runoff model (GR4J) that was coupled with a Snow Accounting Routine (SAR). Results allowed estimation of bulk landslide properties which are typical for strongly damaged rock (porosity 1%, hydraulic conductivity of 1-4·10^-6 m/s). A transient groundwater flow model was then developed to study the impact of the stratified (variably damaged) geometries on recharge, groundwater flow partitioning and pore pressure distribution. We could notably show the importance of state of saturation in the unsaturated zone to allow effective recharge and pore pressure increase at the main sliding surface, especially during snowmelt and summer/fall rainstorms. The pore pressure response to major recharge events ranges from one to 20 days; such variability in pressure diffusion in the vadose zone highlights the importance of the saturation history, typically known for soil slides.